Development of prediction model for hepatocellular carcinoma in chronic hepatitis B patients

Abstract

Abstract Background Chronic hepatitis B infection is one of the leading causes of hepatocellular carcinoma (HCC) worldwide. Despite the endorsement of semi-annual ultrasonography as HCC surveillance method by international societies in at risk population, the accurate prediction of HCC risk is important for public policy strategy in a limited-resource country. Thus, we aim to develop the scoring systems, which are individualized surveillance strategy and cost-effectiveness to assess the risk of HCC in patients with chronic hepatitis B. Methods This retrospective cohort study was conducted to develop a risk estimate model of HCC in chronic hepatitis B (CHB) patients. Our prediction model was derived from data obtained in 2,208 CHB patients from Chulabhorn Hospital, Thailand. (Follow-up period: 2011-2017). Forward stepwise multivariable parametric regression model was applied to obtain coefficients for each predictor. Model input included age, sex, liver cirrhosis, cigarette smoking, alcohol consumption, diabetes mellitus, body mass index, serum HBV DNA level, HBeAg status, alanine aminotransferase, aspartate aminotransferase (AST), alpha-fetoprotein and AST to platelet ratio index. Receiver operating characteristic curves were used to assess discriminatory accuracy of the model. Results During a median follow-up of 6.67 years, 20 cases of HCC were newly diagnosed. Age and liver cirrhosis were statistical significant independent predictors of HCC risk. In the bootstrap simulation (1000 random samplings), the corrected c-index was 0.75 (0.58-0.91). The HCC risk was calculated from the following formula: Age ( Conclusions A simple prediction score constructed from routine clinical and laboratory parameters are accurate in predicting HCC development in Thai patients with CHB infection. Individualized HCC surveillance strategy including surveillance interval and/or alternative surveillance test could be reasonable and cost effectiveness based on our risk scoring. Future prospective validation study is warranted. Legal entity responsible for the study Teerapat Ungtrakul. Funding Chulabhorn Royal Academy. Disclosure All authors have declared no conflicts of interest.